The goal of the proposed project is to validate candidate diagnostic biomarkers in newly developed mouse models for sporadic colon cancer. We have used mass spectrometry to examine the plasma proteomes of mice with a germline Apc mutation to identify a number of tumor-derived proteins that are present at higher levels in the plasma of tumor-bearing mice. In a parallel set of experiments we, along with nine other labs, have identified a set of protein markers that are elevated in humans prior to their eventual clinical diagnosis of colon cancer. From these data we have identified a set of proteins, whose levels of expression in plasma might be used for early detection of colon cancer. As colon cancer is a very heterogeneous disease, it is doubtful if a single protein will be sufficiently robust for all types. To dissect the relationship between individual candidate biomarkers and particular subtypes of colon cancer, genetically engineered mouse (GEM) models can provide an optimal validation vehicle. Traditional mouse models for colon cancer are based on germline modification of carcinogenic genes, resulting in tumors that are mainly localized to the small intestine. Whereas these are reasonable surrogates for familial cancer predisposition syndromes, such as familial adenomatous polyposis, they are poor models for sporadic colon cancer, which makes up 85% of all cases in humans. Furthermore, many of these models present with tumors predominantly in the small intestine. Some even present with extraintestinal tumors. With respect to biomarker validation, these deficiencies need to be addressed. We have developed GEM models that contain floxed Apc and/or Mlh1 genes. When adenovirus expressing cre recombinase is injected into the colon of these mice, isolated tumors present only in the distal colon. We will use antibody-based plasma assays to examine the robustness of our previously identified candidate biomarkers described above. Tumor immunohistochemistry will be used to assess specificity. Longitudinal levels of these proteins in the blood will be correlated with tumor size based on endoscopic metrics. It is possible that this study will yield a panel of robust biomarkers that can be evaluated in human patients for eventual clinical usage. PUBLIC HEALTH RELEVANCE: It is well established that early detection of precancerous lesions leads to greatly improved prognosis. Current colonoscopy surveillance is expensive, uncomfortable, and inconvenient, resulting in poor patient compliance. The development of non-invasive blood protein biomarkers would be of great benefit for patient management and ultimately, overall public health.